Microsoft is “exploring” putting ads in the responses given by Bing Chat, its new search agent powered by OpenAI’s GPT-4. While these sponsored responses are clearly labeled as such, it does make one question how far we’ve really come from the old model of ads on search engines.
Microsoft confirmed this is happening, albeit in an experimental form, in a blog post published today. Here’s the relevant bit from the very end after “a bit of context” explaining no one should be surprised:
We are also exploring additional capabilities for publishers including our more than 7,500 Microsoft Start partner brands. We recently met with some of our partners to begin exploring ideas and to get feedback on how we can continue to distribute content in a way that is meaningful in traffic and revenue for our partners. As we look to continue to evolve the model together, we shared some early ideas we’re exploring including:
• An expanded hover experience where hovering over a link from a publisher will display more links from that publisher giving the user more ways to engage and driving more traffic to the publisher’s website.
• For our Microsoft Start partners, placing a rich caption of Microsoft Start licensed content beside the chat answer helping to drive more user engagement with the content on Microsoft Start where we share the ad revenue with the partner. We’re also exploring placing ads in the chat experience to share the ad revenue with partners whose content contributed to the chat response.
Ads have been reported anecdotally for a couple weeks, but a tweet today by Glean’s Debarghya Das captured them in the wild:
Bing Chat now has Ads!
It's going to be fascinating to see how the unit economics of Ads in language models will unfold and affect search advertising.
1/3 pic.twitter.com/o5YjRjikOP
— Deedy (@deedydas) March 29, 2023
Others pointed out that the ads have been there since launch but not enabled for all users. I tried my best to summon one but to no avail — they don’t seem to be enabled for my account (yet). But clearly this is a sign of things to come.
Let us admit that search engines, and now these search agents, are part of a business model and as such need some kind of monetization. But I think that with all the talk about revolutionizing search and starting fresh, we all rather expected something a little better than sponsored answers.
In the case of the ad above, it’s not clear to the user what’s sponsored here. Is TrueCar selling this car and paying for placement on searches for new Hondas? Or are they just a reference or price checking site that pays to be the preferred provider of car prices to Bing? Why are these prices higher than the ones listed at Honda? Is Microsoft being paid to not include Autotrader or Cars.com listings? Can the user ask for non-sponsored results?
It’s not to say that there is anything nefarious going on. But generally speaking the user should understand what is being advertised. We have learned how to parse search results: ads usually have a little box around them and are at the top of the pile. You don’t have to like it to understand it — and by understanding it you can more intelligently engage (or disengage) with them.
OpenAI releases GPT-4, a multimodal AI that it claims is state-of-the-art
In a case like this, I would love to have a better “meta-conversation,” if you will, a little thought bubble to the side of the chat that has its citations and, if they exist, its ads. The answer itself should not be tampered with — but over in the metacon you have a few links and perhaps a “If you’re thinking of buying a new HR-V, 3 were listed in the last day at TrueCar. (Ad)”
The more dire side effect of this is that ads like this can’t be blocked with current tools. That’s not to say they can’t be blocked at all — it’s not hard to imagine a uBlock Origin that replaces chatbot responses flagged as ads and requests another response. But the paradigm of ads as essentially predictable items in labeled locations is clearly ending.
On one hand, well… good! Everyone hates ads, even though they’re kind of how the web (including, admittedly, this website) pays for itself. On the other, it’s a new and potentially more subtle and subversive form of advertising that may not be as easily delineated and ignored.
Making the user wonder whether the chatbot was paid to say what it said is a great way to slow or erode trust, even with “ad” labels. One can’t help but wonder how honest these companies are being about their deals. Why shouldn’t they have a premium ad tier that influences results but doesn’t get labeled? It’s completely in line with ad industry shenanigans.
While no one expects Microsoft, or Google, Amazon, Meta and all the others to operate these expensive and computation-hungry language models out of the goodness of their hearts (assuming they have hearts and there is good in them), it would be nice to see a little more thought put into how advertising can better be integrated.
When the whole model is changing, the obvious solution — that happens to be a lot like one you used in the old times — is unlikely to be the best. And going with it exposes the company’s priorities in a way that may lead to distrust and skepticism.
Google’s Bard lags behind GPT-4 and Claude in head-to-head comparison
Comment